GAL-2 v0.08 RC: API-Seeded Evidence for Legacy Continuity Across GNU Make and BusyBox

This package documents the strongest current evidence chain for GAL-2 v0.08 core. Scope This release focuses on controlled empirical scenarios involving legacy-style 32-bit consumers operating under an API-seeded GAL-2 continuity model. The package includes three main milestones: Real VM reboot recovery with API-seeded persisted state Sustained repeated GNU Make continuity run BusyBox find -newer discriminator, API-seeded Core evidence chain 1. Real API-seeded authority The strongest v0.08 tests used: real GAL-2 API prime persisted local state local delivery into legacy consumers This package does not treat a fixed-seed local shim as equivalent to full GAL-2. The strongest milestones are interpreted as: real GAL-2 API -> daemon prime -> persisted local state -> local delivery into legacy consumers 2. GNU Make i386 discriminator GNU Make remains a strong real-consumer discriminator. Observed repeatedly: RAW showed visible pathologies such as: Timestamp out of range Clock skew detected a libfaketime-based control remained insufficient GAL-2 preserved correct behavior: Case A -> REBUILD Case B -> up to date 3. Real VM reboot recovery The Debian i386 VM was: primed from the real GAL-2 API rebooted for real recovered from persisted local state re-tested after reboot Observed: persisted state survived the real reboot recovery succeeded after reboot mediated time remained available GNU Make preserved correct GAL-2 behavior after reboot recovery 4. Sustained repeated run A repeated single-node run produced 95 completed iterations in the observed run window. Reviewed first and last iterations showed: stable seed mediated multi_clock_probe correct GNU Make GAL-2 behavior RAW remained pathological libfaketime remained insufficient 5. BusyBox find discriminator, API-seeded BusyBox find -newer was tested after priming from the real GAL-2 API. Observed: RAW failed with Value too large for defined data type the libfaketime-based control also failed with Value too large for defined data type GAL-2 returned the expected results cleanly: Q1 -> ./b.txt and ./c.txt Q2 -> ./c.txt Q3 -> empty GAL-2 exit code was 0 in all three queries This gives the package two real discriminator paths in the tested evidence chain: GNU Make i386 BusyBox find Strongest defensible claim At this stage, GAL-2 v0.08 core has demonstrated two real discriminator paths in the tested evidence chain, both under an API-seeded continuity model: GNU Make i386 BusyBox find In the tested scenarios, API-seeded GAL-2 preserved correct behavior where RAW execution remained pathological or failed and a libfaketime-based control remained insufficient. Scope note This is not a universal claim for all kernels, all libc stacks, or all legacy binaries. It is a controlled empirical result for the included scenarios. Important limitations Not every Unix tool discriminated in every exploratory scenario. Some exploratory multi-binary probes remain exploratory rather than load-bearing evidence. The deeper internal API metadata fields are not yet publicly demonstrated as functionally consumed by the local path in this package. Package integrity Final package SHA256: 644b92eb370577ed3308e566f77e3084d6d7c8cf472a41e17d268c4f5486a6a5